Skip to content

crypto/internal/bigmod: flag amd64 assembly as noescape #58501

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom
Closed
FiloSottile opened this issue Feb 13, 2023 · 4 comments
Closed

crypto/internal/bigmod: flag amd64 assembly as noescape #58501

FiloSottile opened this issue Feb 13, 2023 · 4 comments
Labels
FrozenDueToAge NeedsFix The path to resolution is known, but the work has not been done. Performance
Milestone
Go1.21

Comments

@FiloSottile
Copy link
Contributor

I had overlooked it, causing a performance loss.

name                    old time/op    new time/op    delta
DecryptPKCS1v15/2048-8    1.50ms ± 0%    1.48ms ± 0%   -0.95%  (p=0.000 n=9+10)
DecryptPKCS1v15/3072-8    4.64ms ± 1%    4.60ms ± 0%   -0.82%  (p=0.000 n=8+10)
DecryptPKCS1v15/4096-8    10.7ms ± 0%    10.6ms ± 1%   -0.99%  (p=0.000 n=10+10)
EncryptPKCS1v15/2048-8     158µs ± 0%     157µs ± 0%   -0.63%  (p=0.000 n=10+10)
DecryptOAEP/2048-8        1.50ms ± 0%    1.48ms ± 0%   -1.09%  (p=0.000 n=9+10)
EncryptOAEP/2048-8         161µs ± 0%     160µs ± 0%   -0.34%  (p=0.000 n=9+10)
SignPKCS1v15/2048-8       1.55ms ± 0%    1.53ms ± 1%   -1.32%  (p=0.000 n=10+10)
VerifyPKCS1v15/2048-8      157µs ± 0%     157µs ± 0%   -0.33%  (p=0.004 n=9+10)
SignPSS/2048-8            1.55ms ± 0%    1.54ms ± 0%   -1.14%  (p=0.000 n=10+10)
VerifyPSS/2048-8           160µs ± 0%     160µs ± 0%   -0.32%  (p=0.000 n=10+10)

name                    old alloc/op   new alloc/op   delta
DecryptPKCS1v15/2048-8    15.0kB ± 0%     0.6kB ± 0%  -95.74%  (p=0.000 n=10+10)
DecryptPKCS1v15/3072-8    17.9kB ± 0%     3.5kB ± 0%  -80.65%  (p=0.000 n=10+10)
DecryptPKCS1v15/4096-8    19.1kB ± 0%     4.7kB ± 0%  -75.25%  (p=0.000 n=10+10)
EncryptPKCS1v15/2048-8    7.51kB ± 0%    1.17kB ± 0%  -84.39%  (p=0.000 n=10+10)
DecryptOAEP/2048-8        15.3kB ± 0%     0.9kB ± 0%  -94.29%  (p=0.000 n=10+10)
EncryptOAEP/2048-8        7.74kB ± 0%    1.40kB ± 0%  -81.86%  (p=0.000 n=10+10)
SignPKCS1v15/2048-8       21.6kB ± 0%     0.9kB ± 0%  -95.86%  (p=0.000 n=10+10)
VerifyPKCS1v15/2048-8     7.25kB ± 0%    0.91kB ± 0%  -87.42%  (p=0.000 n=10+10)
SignPSS/2048-8            22.0kB ± 0%     1.3kB ± 0%  -94.12%  (p=0.000 n=10+10)
VerifyPSS/2048-8          7.46kB ± 0%    1.12kB ± 0%  -84.98%  (p=0.000 n=10+10)

name                    old allocs/op  new allocs/op  delta
DecryptPKCS1v15/2048-8      54.0 ± 0%       4.0 ± 0%  -92.59%  (p=0.000 n=10+10)
DecryptPKCS1v15/3072-8      60.0 ± 0%      10.0 ± 0%  -83.33%  (p=0.000 n=10+10)
DecryptPKCS1v15/4096-8      60.0 ± 0%      10.0 ± 0%  -83.33%  (p=0.000 n=10+10)
EncryptPKCS1v15/2048-8      29.0 ± 0%       7.0 ± 0%  -75.86%  (p=0.000 n=10+10)
DecryptOAEP/2048-8          60.0 ± 0%      10.0 ± 0%  -83.33%  (p=0.000 n=10+10)
EncryptOAEP/2048-8          35.0 ± 0%      13.0 ± 0%  -62.86%  (p=0.000 n=10+10)
SignPKCS1v15/2048-8         77.0 ± 0%       5.0 ± 0%  -93.51%  (p=0.000 n=10+10)
VerifyPKCS1v15/2048-8       28.0 ± 0%       6.0 ± 0%  -78.57%  (p=0.000 n=10+10)
SignPSS/2048-8              82.0 ± 0%      10.0 ± 0%  -87.80%  (p=0.000 n=10+10)
VerifyPSS/2048-8            33.0 ± 0%      11.0 ± 0%  -66.67%  (p=0.000 n=10+10)
@gopherbot
Copy link
Contributor

Change https://go.dev/cl/460542 mentions this issue: crypto/internal/bigmod: flag amd64 assembly as noescape

@FiloSottile
Copy link
Contributor Author

@gopherbot please open a Go 1.20 backport. This is arguably partially mitigating a performance regression in Go 1.20, it's a "bug" in new code introduced in Go 1.20, and the patch is as safe as it gets.

@gopherbot gopherbot mentioned this issue Feb 13, 2023
Closed
@gopherbot
Copy link
Contributor

Backport issue(s) opened: #58505 (for 1.20).

Remember to create the cherry-pick CL(s) as soon as the patch is submitted to master, according to https://go.dev/wiki/MinorReleases.

@prattmic prattmic added this to the Go1.21 milestone Feb 14, 2023
@prattmic prattmic added the NeedsFix The path to resolution is known, but the work has not been done. label Feb 14, 2023
johanbrandhorst pushed a commit to Pryz/go that referenced this issue Feb 22, 2023
@FiloSottile @johanbrandhorst
crypto/internal/bigmod: flag amd64 assembly as noescape
9184550 
I had forgotten, which caused amd64 allocations to go back up
significantly. Added an allocations test.

name                    old time/op    new time/op    delta
DecryptPKCS1v15/2048-8    1.50ms ± 0%    1.48ms ± 0%   -0.95%  (p=0.000 n=9+10)
DecryptPKCS1v15/3072-8    4.64ms ± 1%    4.60ms ± 0%   -0.82%  (p=0.000 n=8+10)
DecryptPKCS1v15/4096-8    10.7ms ± 0%    10.6ms ± 1%   -0.99%  (p=0.000 n=10+10)
EncryptPKCS1v15/2048-8     158µs ± 0%     157µs ± 0%   -0.63%  (p=0.000 n=10+10)
DecryptOAEP/2048-8        1.50ms ± 0%    1.48ms ± 0%   -1.09%  (p=0.000 n=9+10)
EncryptOAEP/2048-8         161µs ± 0%     160µs ± 0%   -0.34%  (p=0.000 n=9+10)
SignPKCS1v15/2048-8       1.55ms ± 0%    1.53ms ± 1%   -1.32%  (p=0.000 n=10+10)
VerifyPKCS1v15/2048-8      157µs ± 0%     157µs ± 0%   -0.33%  (p=0.004 n=9+10)
SignPSS/2048-8            1.55ms ± 0%    1.54ms ± 0%   -1.14%  (p=0.000 n=10+10)
VerifyPSS/2048-8           160µs ± 0%     160µs ± 0%   -0.32%  (p=0.000 n=10+10)

name                    old alloc/op   new alloc/op   delta
DecryptPKCS1v15/2048-8    15.0kB ± 0%     0.6kB ± 0%  -95.74%  (p=0.000 n=10+10)
DecryptPKCS1v15/3072-8    17.9kB ± 0%     3.5kB ± 0%  -80.65%  (p=0.000 n=10+10)
DecryptPKCS1v15/4096-8    19.1kB ± 0%     4.7kB ± 0%  -75.25%  (p=0.000 n=10+10)
EncryptPKCS1v15/2048-8    7.51kB ± 0%    1.17kB ± 0%  -84.39%  (p=0.000 n=10+10)
DecryptOAEP/2048-8        15.3kB ± 0%     0.9kB ± 0%  -94.29%  (p=0.000 n=10+10)
EncryptOAEP/2048-8        7.74kB ± 0%    1.40kB ± 0%  -81.86%  (p=0.000 n=10+10)
SignPKCS1v15/2048-8       21.6kB ± 0%     0.9kB ± 0%  -95.86%  (p=0.000 n=10+10)
VerifyPKCS1v15/2048-8     7.25kB ± 0%    0.91kB ± 0%  -87.42%  (p=0.000 n=10+10)
SignPSS/2048-8            22.0kB ± 0%     1.3kB ± 0%  -94.12%  (p=0.000 n=10+10)
VerifyPSS/2048-8          7.46kB ± 0%    1.12kB ± 0%  -84.98%  (p=0.000 n=10+10)

name                    old allocs/op  new allocs/op  delta
DecryptPKCS1v15/2048-8      54.0 ± 0%       4.0 ± 0%  -92.59%  (p=0.000 n=10+10)
DecryptPKCS1v15/3072-8      60.0 ± 0%      10.0 ± 0%  -83.33%  (p=0.000 n=10+10)
DecryptPKCS1v15/4096-8      60.0 ± 0%      10.0 ± 0%  -83.33%  (p=0.000 n=10+10)
EncryptPKCS1v15/2048-8      29.0 ± 0%       7.0 ± 0%  -75.86%  (p=0.000 n=10+10)
DecryptOAEP/2048-8          60.0 ± 0%      10.0 ± 0%  -83.33%  (p=0.000 n=10+10)
EncryptOAEP/2048-8          35.0 ± 0%      13.0 ± 0%  -62.86%  (p=0.000 n=10+10)
SignPKCS1v15/2048-8         77.0 ± 0%       5.0 ± 0%  -93.51%  (p=0.000 n=10+10)
VerifyPKCS1v15/2048-8       28.0 ± 0%       6.0 ± 0%  -78.57%  (p=0.000 n=10+10)
SignPSS/2048-8              82.0 ± 0%      10.0 ± 0%  -87.80%  (p=0.000 n=10+10)
VerifyPSS/2048-8            33.0 ± 0%      11.0 ± 0%  -66.67%  (p=0.000 n=10+10)

Fixes golang#58501

Change-Id: I418c5152833787b80220b556336ec284674c2493
Reviewed-on: https://go-review.googlesource.com/c/go/+/460542
Run-TryBot: Filippo Valsorda <filippo@golang.org>
TryBot-Result: Gopher Robot <gobot@golang.org>
Reviewed-by: Roland Shoemaker <roland@golang.org>
Reviewed-by: Michael Pratt <mpratt@google.com>
Auto-Submit: Filippo Valsorda <filippo@golang.org>
@gopherbot
Copy link
Contributor

Change https://go.dev/cl/471855 mentions this issue: [release-branch.go1.20] crypto/internal/bigmod: flag amd64 assembly as noescape

gopherbot pushed a commit that referenced this issue Feb 28, 2023
@FiloSottile @gopherbot
[release-branch.go1.20] crypto/internal/bigmod: flag amd64 assembly a…
ef79380 
…s noescape

I had forgotten, which caused amd64 allocations to go back up
significantly. Added an allocations test.

name                    old time/op    new time/op    delta
DecryptPKCS1v15/2048-8    1.50ms ± 0%    1.48ms ± 0%   -0.95%  (p=0.000 n=9+10)
DecryptPKCS1v15/3072-8    4.64ms ± 1%    4.60ms ± 0%   -0.82%  (p=0.000 n=8+10)
DecryptPKCS1v15/4096-8    10.7ms ± 0%    10.6ms ± 1%   -0.99%  (p=0.000 n=10+10)
EncryptPKCS1v15/2048-8     158µs ± 0%     157µs ± 0%   -0.63%  (p=0.000 n=10+10)
DecryptOAEP/2048-8        1.50ms ± 0%    1.48ms ± 0%   -1.09%  (p=0.000 n=9+10)
EncryptOAEP/2048-8         161µs ± 0%     160µs ± 0%   -0.34%  (p=0.000 n=9+10)
SignPKCS1v15/2048-8       1.55ms ± 0%    1.53ms ± 1%   -1.32%  (p=0.000 n=10+10)
VerifyPKCS1v15/2048-8      157µs ± 0%     157µs ± 0%   -0.33%  (p=0.004 n=9+10)
SignPSS/2048-8            1.55ms ± 0%    1.54ms ± 0%   -1.14%  (p=0.000 n=10+10)
VerifyPSS/2048-8           160µs ± 0%     160µs ± 0%   -0.32%  (p=0.000 n=10+10)

name                    old alloc/op   new alloc/op   delta
DecryptPKCS1v15/2048-8    15.0kB ± 0%     0.6kB ± 0%  -95.74%  (p=0.000 n=10+10)
DecryptPKCS1v15/3072-8    17.9kB ± 0%     3.5kB ± 0%  -80.65%  (p=0.000 n=10+10)
DecryptPKCS1v15/4096-8    19.1kB ± 0%     4.7kB ± 0%  -75.25%  (p=0.000 n=10+10)
EncryptPKCS1v15/2048-8    7.51kB ± 0%    1.17kB ± 0%  -84.39%  (p=0.000 n=10+10)
DecryptOAEP/2048-8        15.3kB ± 0%     0.9kB ± 0%  -94.29%  (p=0.000 n=10+10)
EncryptOAEP/2048-8        7.74kB ± 0%    1.40kB ± 0%  -81.86%  (p=0.000 n=10+10)
SignPKCS1v15/2048-8       21.6kB ± 0%     0.9kB ± 0%  -95.86%  (p=0.000 n=10+10)
VerifyPKCS1v15/2048-8     7.25kB ± 0%    0.91kB ± 0%  -87.42%  (p=0.000 n=10+10)
SignPSS/2048-8            22.0kB ± 0%     1.3kB ± 0%  -94.12%  (p=0.000 n=10+10)
VerifyPSS/2048-8          7.46kB ± 0%    1.12kB ± 0%  -84.98%  (p=0.000 n=10+10)

name                    old allocs/op  new allocs/op  delta
DecryptPKCS1v15/2048-8      54.0 ± 0%       4.0 ± 0%  -92.59%  (p=0.000 n=10+10)
DecryptPKCS1v15/3072-8      60.0 ± 0%      10.0 ± 0%  -83.33%  (p=0.000 n=10+10)
DecryptPKCS1v15/4096-8      60.0 ± 0%      10.0 ± 0%  -83.33%  (p=0.000 n=10+10)
EncryptPKCS1v15/2048-8      29.0 ± 0%       7.0 ± 0%  -75.86%  (p=0.000 n=10+10)
DecryptOAEP/2048-8          60.0 ± 0%      10.0 ± 0%  -83.33%  (p=0.000 n=10+10)
EncryptOAEP/2048-8          35.0 ± 0%      13.0 ± 0%  -62.86%  (p=0.000 n=10+10)
SignPKCS1v15/2048-8         77.0 ± 0%       5.0 ± 0%  -93.51%  (p=0.000 n=10+10)
VerifyPKCS1v15/2048-8       28.0 ± 0%       6.0 ± 0%  -78.57%  (p=0.000 n=10+10)
SignPSS/2048-8              82.0 ± 0%      10.0 ± 0%  -87.80%  (p=0.000 n=10+10)
VerifyPSS/2048-8            33.0 ± 0%      11.0 ± 0%  -66.67%  (p=0.000 n=10+10)

Updates #58501.
Fixes #58505.

Change-Id: I418c5152833787b80220b556336ec284674c2493
Reviewed-on: https://go-review.googlesource.com/c/go/+/460542
Run-TryBot: Filippo Valsorda <filippo@golang.org>
TryBot-Result: Gopher Robot <gobot@golang.org>
Reviewed-by: Roland Shoemaker <roland@golang.org>
Reviewed-by: Michael Pratt <mpratt@google.com>
Auto-Submit: Filippo Valsorda <filippo@golang.org>
(cherry picked from commit ed370d8)
Reviewed-on: https://go-review.googlesource.com/c/go/+/471855
Auto-Submit: Dmitri Shuralyov <dmitshur@google.com>
Run-TryBot: Dmitri Shuralyov <dmitshur@golang.org>
Reviewed-by: Dmitri Shuralyov <dmitshur@google.com>
romaindoumenc pushed a commit to TroutSoftware/go that referenced this issue Mar 3, 2023
@FiloSottile @romaindoumenc
[release-branch.go1.20] crypto/internal/bigmod: flag amd64 assembly a…
c64f81e 
…s noescape

I had forgotten, which caused amd64 allocations to go back up
significantly. Added an allocations test.

name                    old time/op    new time/op    delta
DecryptPKCS1v15/2048-8    1.50ms ± 0%    1.48ms ± 0%   -0.95%  (p=0.000 n=9+10)
DecryptPKCS1v15/3072-8    4.64ms ± 1%    4.60ms ± 0%   -0.82%  (p=0.000 n=8+10)
DecryptPKCS1v15/4096-8    10.7ms ± 0%    10.6ms ± 1%   -0.99%  (p=0.000 n=10+10)
EncryptPKCS1v15/2048-8     158µs ± 0%     157µs ± 0%   -0.63%  (p=0.000 n=10+10)
DecryptOAEP/2048-8        1.50ms ± 0%    1.48ms ± 0%   -1.09%  (p=0.000 n=9+10)
EncryptOAEP/2048-8         161µs ± 0%     160µs ± 0%   -0.34%  (p=0.000 n=9+10)
SignPKCS1v15/2048-8       1.55ms ± 0%    1.53ms ± 1%   -1.32%  (p=0.000 n=10+10)
VerifyPKCS1v15/2048-8      157µs ± 0%     157µs ± 0%   -0.33%  (p=0.004 n=9+10)
SignPSS/2048-8            1.55ms ± 0%    1.54ms ± 0%   -1.14%  (p=0.000 n=10+10)
VerifyPSS/2048-8           160µs ± 0%     160µs ± 0%   -0.32%  (p=0.000 n=10+10)

name                    old alloc/op   new alloc/op   delta
DecryptPKCS1v15/2048-8    15.0kB ± 0%     0.6kB ± 0%  -95.74%  (p=0.000 n=10+10)
DecryptPKCS1v15/3072-8    17.9kB ± 0%     3.5kB ± 0%  -80.65%  (p=0.000 n=10+10)
DecryptPKCS1v15/4096-8    19.1kB ± 0%     4.7kB ± 0%  -75.25%  (p=0.000 n=10+10)
EncryptPKCS1v15/2048-8    7.51kB ± 0%    1.17kB ± 0%  -84.39%  (p=0.000 n=10+10)
DecryptOAEP/2048-8        15.3kB ± 0%     0.9kB ± 0%  -94.29%  (p=0.000 n=10+10)
EncryptOAEP/2048-8        7.74kB ± 0%    1.40kB ± 0%  -81.86%  (p=0.000 n=10+10)
SignPKCS1v15/2048-8       21.6kB ± 0%     0.9kB ± 0%  -95.86%  (p=0.000 n=10+10)
VerifyPKCS1v15/2048-8     7.25kB ± 0%    0.91kB ± 0%  -87.42%  (p=0.000 n=10+10)
SignPSS/2048-8            22.0kB ± 0%     1.3kB ± 0%  -94.12%  (p=0.000 n=10+10)
VerifyPSS/2048-8          7.46kB ± 0%    1.12kB ± 0%  -84.98%  (p=0.000 n=10+10)

name                    old allocs/op  new allocs/op  delta
DecryptPKCS1v15/2048-8      54.0 ± 0%       4.0 ± 0%  -92.59%  (p=0.000 n=10+10)
DecryptPKCS1v15/3072-8      60.0 ± 0%      10.0 ± 0%  -83.33%  (p=0.000 n=10+10)
DecryptPKCS1v15/4096-8      60.0 ± 0%      10.0 ± 0%  -83.33%  (p=0.000 n=10+10)
EncryptPKCS1v15/2048-8      29.0 ± 0%       7.0 ± 0%  -75.86%  (p=0.000 n=10+10)
DecryptOAEP/2048-8          60.0 ± 0%      10.0 ± 0%  -83.33%  (p=0.000 n=10+10)
EncryptOAEP/2048-8          35.0 ± 0%      13.0 ± 0%  -62.86%  (p=0.000 n=10+10)
SignPKCS1v15/2048-8         77.0 ± 0%       5.0 ± 0%  -93.51%  (p=0.000 n=10+10)
VerifyPKCS1v15/2048-8       28.0 ± 0%       6.0 ± 0%  -78.57%  (p=0.000 n=10+10)
SignPSS/2048-8              82.0 ± 0%      10.0 ± 0%  -87.80%  (p=0.000 n=10+10)
VerifyPSS/2048-8            33.0 ± 0%      11.0 ± 0%  -66.67%  (p=0.000 n=10+10)

Updates golang#58501.
Fixes golang#58505.

Change-Id: I418c5152833787b80220b556336ec284674c2493
Reviewed-on: https://go-review.googlesource.com/c/go/+/460542
Run-TryBot: Filippo Valsorda <filippo@golang.org>
TryBot-Result: Gopher Robot <gobot@golang.org>
Reviewed-by: Roland Shoemaker <roland@golang.org>
Reviewed-by: Michael Pratt <mpratt@google.com>
Auto-Submit: Filippo Valsorda <filippo@golang.org>
(cherry picked from commit ed370d8)
Reviewed-on: https://go-review.googlesource.com/c/go/+/471855
Auto-Submit: Dmitri Shuralyov <dmitshur@google.com>
Run-TryBot: Dmitri Shuralyov <dmitshur@golang.org>
Reviewed-by: Dmitri Shuralyov <dmitshur@google.com>
@golang golang locked and limited conversation to collaborators Feb 28, 2024
Sign up for free to subscribe to this conversation on GitHub. Already have an account? Sign in.
Assignees
No one assigned
Labels
FrozenDueToAge NeedsFix The path to resolution is known, but the work has not been done. Performance
Projects
None yet
Milestone
Go1.21
Development

No branches or pull requests
4 participants
@FiloSottile @prattmic @dmitshur @gopherbot